Connectome-Based Predictive Modeling of Individual Anxiety

Area of Science:

• Neuroscience
• Psychiatry
• Computational Biology

Background:

• Anxiety-related illnesses are common, necessitating biomarkers for early risk assessment.
• Identifying neurobiological markers for trait anxiety can improve mental illness risk evaluation.

Purpose of the Study:

• To predict individual trait anxiety levels using whole-brain resting-state functional connectivity (rsFC).
• To investigate the contribution of specific brain regions and their connections to anxiety prediction using connectome-based predictive modeling (CPM).

Main Methods:

• Applied CPM to rsFC data from 76 healthy participants to predict trait anxiety.
• Utilized a computational 'lesion' approach to assess the importance of brain areas and their connectivity.
• Validated the predictive model on an independent sample of 501 individuals.

Main Results:

• CPM successfully predicted individual trait anxiety based on whole-brain rsFC.
• Connectivity between limbic areas and the prefrontal cortex was particularly important for prediction.
• The model's predictive power significantly decreased after simulating lesions in limbic areas and their connections to the prefrontal cortex.

Conclusions:

• Limbic system and prefrontal cortex connectivity play critical roles in predicting trait anxiety.
• Connectome-based modeling is a valuable tool for predicting individual personality differences.
• This approach shows potential for identifying individuals at risk for psychopathology based on personality factors.